Biological membrane proteins participate in a wide variety of vital cellular functions such as material exchange, energy conservation, cellular communication, and detoxification. As estimated by the pharmaceutical industries, 95% of all available drugs interact with targets that are membrane bound. In the human genome, membrane proteins constitute 30% of open reading frames, yet the number of membrane protein structures determined to date is disproportionally small. Although structural knowledge for this class of proteins is being actively pursued, even more so following the initiative of the NIH road map, structural data for membrane proteins at atomic resolution are only being obtained slowly due to various practical obstacles. The picture gets even worse for eukaryotic membrane proteins; excluding those of mitochondrial origin, only a few eukaryotic membrane protein structures have been determined to date at atomic resolution. My group studies the structure and function of a few selected families of membrane proteins using a combination of biochemical and crystallographic methods. One protein family of particular interest includes those involved in cellular multidrug resistance such as P-glycoprotein (P-gp) and related ABC family proteins. In collaboration with Drs. Ambudkar and Gottesman of LCB, we are actively pursuing crystallization of P-gp. In addition, my lab has been working on expression, purification and crystallization of various P-gp related proteins for crystallographic studies. Technically, to obtain a membrane protein structure, four major obstacles must be overcome: (1) achieving high-level protein expression, (2) obtaining pure and mono-dispersed proteins in large quantity, (3) growing diffraction quality crystals, and (4) solving crystallographic phase problems, often at relatively low resolutions, for membrane proteins. All these difficulties are due in large part to the fact that membrane proteins have large hydrophobic surfaces. Membrane proteins, especially eukaryotic membrane proteins, are difficult to express in large quantities and in active forms in commonly available expression systems. Successful heterologous expressions of membrane proteins are always achieved in a case-by-case situation. Currently, the most common approach to express large amounts of active membrane protein is to screen for high-level expression of a large number of homologs, mostly those of prokaryotes in various expression systems. Achieving high-level expression of membrane protein is perhaps one or two orders of magnitude more difficult than high-level expression of soluble proteins. Even in the case of very high-level expression, the amount of the target membrane protein is rarely over 30% of the total membrane proteins in cell membrane (2% for P-gp, 5% for LmrA, and 2% for Pdr5p), making purification of a large amount of protein difficult without large-scale fermentation and cell disruption facilities. Furthermore, successful high-level expression in test tubes does not guarantee success when scaling up. Membrane proteins are often associated with each other in a non-specific manner (polydispersity) when purified, which is detrimental to successful crystallization; this problem can sometimes be eliminated by screening for different detergents and solvent conditions. When purified, membrane proteins exist in solution as protein-detergent complexes; the available hydrophilic surface that is useful for specific crystal contact is limited. It is quite common to screen for over 30,000 conditions before a diffraction quality crystal form can be found. It has been successful in a few cases to artificially increase the hydrophilic surface by attaching conformational sensitive monoclonal antibodies to target membrane proteins. Our group is also working towards improving existing or devising new technologies to facilitate crystallization of membrane proteins.